Abstract: A disclosed system, method and computer readable storage medium is provided to receive an input video comprising a plurality of video frames, each of which has a plurality of input pixels. Each input pixel is of n-bit depth (e.g., 8-bit depth). The method optionally increases the bitdepth of each input pixel by a predetermined factor (e.g., 4-bit factor). The method further compresses and reconstructs each pixels of the input video. The method further compresses and reconstructs each pixel of intermediate frame buffers stored in a decoder frame buffer. The method calculates an adaptive offset for each block (e.g., of size of 4×4) of reconstructed pixel values of the intermediate frame buffers. Furthermore, the method can compute two-sub-block offset values and uses them to compute an optimized adaptive offset to be applied to each pixel within the corresponding sub-block.

Abstract: A system (and a method) are disclosed for reducing intra frame beating effect at frame level and macroblock (MB) level within a video processing system. The frame-level system includes a pre-encoding unit, a reconstruction unit, an intra-beat frame evaluation unit, an intra-beat frame selection unit and an entropy encoding unit. The pre-encoding unit identifies an intra-beat frame of current group of pictures (GOP) and its corresponding intra-beat reference frame, and predicts the intra-beat frame by its intra-beat reference frame. The intra-beat frame evaluation unit calculates an intra-beat frame score. The intra-beat frame selection unit selects an intra-beat frame for re-encoding. The macroblock-level system includes corresponding processing units as frame-level system, but configured to efficiently process MBs. The macroblock-level system allows each MB of an intra-beat frame to be selectively pre-filtered using its intra-beat reference frame.

Abstract: An apparatus and method for generating predictors performs motion estimation of a target macroblock in a target field against data segments in reference fields. The same motion estimation engine is used to perform various image processing operations to efficiently use resources of the apparatus. Different reference fields are used depending on modes of operation. In a deinterlacing mode, deinterlacing is performed using directional interpolation, recursive motion compensated deinterlacing, and motion adaptive deinterlacing.

Abstract: An apparatus and method for generating predictors performs motion estimation of a target macroblock in a target field against data segments in reference fields. The same motion estimation engine is used to perform various image processing operations to efficiently use resources of the apparatus. Different reference fields are used depending on modes of operation. In a deinterlacing mode, deinterlacing is performed using directional interpolation, recursive motion compensated deinterlacing, and motion adaptive deinterlacing.

Abstract: An apparatus and method for generating predictors performs motion estimation of a target macroblock in a target field against data segments in reference fields. The same motion estimation engine is used to perform various image processing operations to efficiently use resources of the apparatus. Different reference fields are used depending on modes of operation. In a deinterlacing mode, deinterlacing is performed using directional interpolation, recursive motion compensated deinterlacing, and motion adaptive deinterlacing.

Abstract: A system (and a method) for compressing reference frames in a video transcoder. A transcoder receives a compressed input stream in a first compressed format and output a compressed output stream in a second compressed format. A decoder and an encoder in the transcoder use compressed reference frames. The reference frames are compressed by transforming a block of pixels from a spatial domain to a frequency domain to generate a coefficient array. The coefficient array is quantized and encoded to compress the size of the coefficients array to the size of a fixed bucket. The values of the entropy coded and quantized array are stored in a memory for use in decoding and/or encoding.

Abstract: A system (and a method) for compressing reference frames in a video transcoder. A transcoder receives a compressed input stream in a first compressed format and output a compressed output stream in a second compressed format. A decoder and an encoder in the transcoder use compressed reference frames. The reference frames are compressed by transforming a block of pixels from a spatial domain to a frequency domain to generate a coefficient array. The coefficient array is quantized and encoded to compress the size of the coefficients array to the size of a fixed bucket. The values of the entropy coded and quantized array are stored in a memory for use in decoding and/or encoding.

Abstract: One embodiment of the present invention provides a method and system for transforming a video bitstream in an interlaced format into a progressive format which can be displayed by a digital television. For example, the present embodiment utilizes the pixel information of a current field, previous field, and future field of the interlaced video bitstream to try to determine what the original content is of the missing lines of the current field. Specifically, the present embodiment utilizes different sets of pixel information in order to estimate the amount of motion that exist within a video bitstream. In this manner, the present embodiment is able to more closely determine the original value of the missing pixels of each field of the interlaced video bitstream. Therefore, the present embodiment provides a de-interlacing function enabling digital televisions to receive interlaced video bitstreams and display them in the progressive format.

Abstract: One embodiment of the present invention provides a method and system for transforming a video bitstream in an interlaced format into a progressive format which can be displayed by a digital television. For example, the present embodiment utilizes the pixel information of a current field, previous field, and future field of the interlaced video bitstream to try to determine what the original content is of the missing lines of the current field. Specifically, the present embodiment utilizes different sets of pixel information in order to estimate the amount of motion that exist within a video bitstream. In this manner, the present embodiment is able to more closely determine the original value of the missing pixels of each field of the interlaced video bitstream. Therefore, the present embodiment provides a de-interlacing function enabling digital televisions to receive interlaced video bitstreams and display them in the progressive format.

Abstract: A video decoder for decoding compressed high-definition video data. The video decoder includes a down-converter and an up-converter placed in the decoder's data path to reduce the amount of video data that needs to be stored in frame buffer memory. Reference frame video data is decimated and low-pass filtered by the down-converter for storage in a frame buffer memory. The up-converter uses interpolation and low-pass with high boost filtering to reconstruct the reference frames on-the-fly for use in motion compensated prediction.

Abstract: A video decoder for decoding compressed high-definition video data. The video decoder includes compress and decompress circuits placed in the decoder's datapath to reduce the amount of video data that needs to be stored in frame buffer memory. Reference frame video data is compressed by the compress circuit for storage in frame buffer memory. A rate control circuit is provided to control the rate at which the compress circuit stores reference frame video data in the frame buffer memory. The decompress circuit reconstructs the reference frames on-the-fly for use in motion compensated prediction. Methods of decoding compressed high-definition video signals are also disclosed.

Abstract: Systems and methods are provided for allocating bits to pictures in accordance with the bit allocation constraints for operation at both constant and variable bit rates. The encoding complexity of each picture is modeled based on its spatial and temporal properties. Based on the modeling, a number of bits is allocated in advance to each picture. Each picture is then compressed according to the bits allocated. Based on the difference between the achieved and desired bit rates and on the availability of computational resources, a second pass or multiple passes may be performed to improve the coding.